The Golden Dark Age Of the Internet
Before video; engagement, content aggregators, and the algorithm
For those of you whose experience in the internet is relatively modern - there was a day before the modern search engine, before video, before algorithmic recommendations.
When the function that is now performed by the algorithm was an artisanal craft; when there were sites you would go to whose sole purpose was to direct you to interesting new content.
The fact that you're reading this puts you in an odd camp; the people whose relation to the internet still includes longform text. Or perhaps I should say “now” includes longform text. Maybe, like me, you find text inherently more interesting. Maybe it is just part of a more varied diet which includes many forms of media. Maybe this is just an odd fluke, I couldn't say.
But at this point, longform text conversations on the internet have become so old they're new again; just as Discord has effectively revived chat rooms and instant messages, platforms like the one I’m writing this on are part of a revival of blogs, which for about a decade there, were fading away. There are some dinosaurs out there - many have changed outlets, for reasons ranging from monetization to fear of censorship. But others are still plugging away at something they’ve been doing for ages.
And yet most of them are gone; many deleted. There was a blog I read many, many years ago, authored by a gentleman (I presume) who went by the monikor of “8 ball”; his favorite topic seemed to have been the practice of witchcraft in Africa, but he covered many others as well. My favorite post exists, as far as I know, only in memory, about Noah’s Ark, with a specific focus on the logistics that would have been necessary, and discussing what exactly the units of measurement would translate to, and how large the boat would need to be to hold “two of every kind”. (It concluded that Noah -caused- the flood, by pushing a city-sized boat into the sea after cutting down a forest to build it. I do not do it justice, here; he was funnier than I am.)
Alas, I believe he deleted the blog, and to my knowledge this deletion predates anything like the internet archive. But I had stopped reading him at that point; he was one of many comedians who fell prey to what is now sometimes referred to as Bush Derangement Syndrome, and stopped being funny (well, at least to me - I assume somebody found the unhinged anti-Bush rants to be amusing). “The internet is forever” wasn’t - and isn’t - true. Things just … disappear.
So if you don’t have the context, this is, in a significant way, a return to the golden age of the Internet. In another, we’re still in the internet dark ages, as companies have struggled to find a balance in the political knife-edge we have found ourselves in over the last decade.
Sometimes people talk about how, no, it’s always been like this. And to some extent this is true. But to a greater extent - no, the last decade truly has been exceptionally partisan. Some would blame the right, some would blame the left; arguments abound over who fired first. (Remember when The Dixie Chicks got canceled for criticizing the Bush administration? No?) But the question of who fired first is kind of meaningless; if you’re looking for when the other guy “fired first”, you can find your answer. And the other guy can find an answer that predates yours.
Really, the question is who is going to stop. And I get the optimistic impression that the answer might be “Everyone, because we’re tired of this shit.” There is, I think, political room for somebody who doesn’t take every opportunity to snipe at the other side, a tendency that has been slowly escalating since at least LBJ. (Did he actually fire first? Silly question.)
When I was younger, and a little bit more naive, I thought the problem was that we had no common enemy to unite against since the fall of the Soviet Union. But no, even the unity we possessed then was illusory, enforced by widespread censorship of the media.
No, the problem is, basically, that it has historically been fun to be a little bit naughty, to break the rules just a little bit; it was delightful when somebody skewered your enemy with a well-aimed shot. But it’s just not fun anymore; people are exhausted of this kind of thing. We could say we’re more enlightened now, but I don’t think that’s accurate; rather, it’s just gotten boring and predictable.
What changed?
It stopped being naughty, is the short of it, and just became routine. Everybody does it now, and, having lost its novelty, it stopped being funny. Even The Daily Show stopped making fun of Republicans, for just a little while, after 9/11; today, such behavior is so routine that to stop for something like that is utterly alien.
We’re not out of the dark ages yet. But we do seem to be going into something like a new golden age, tempered by the past two decades of experience, and now armed with the knowledge that a platform is an ephemeral thing, subject to being taken away at a whim.
And it is a bigger internet; video and audio have joined text.
Perhaps soon we’ll have a platform for content curation; subscribe to my channel, and I’ll provide you an endless series of fascinating things to read, or listen to, or watch! You can see microcosms swirling in many places, alongside algorithmic recommendations. Video seems to be one of the odd holdouts there, but perhaps that’s just my own preference for text preventing me from encountering them. There are still a few dinosaurs of content curation trundling along, of course, such as the vaguely right-libertarian Instapundit, which has been a useful tool for me, personally, to reach outside my left-wing social bubble and see what other people are thinking.
For those of you also in left-wing social bubbles, who may have encountered the sporadic but consistent stream of “pedophile priests” or “pedophile Republican” narratives - the right wing has its own sporadic but consistent stream of “pedophile teachers” and “pedophile Democrats”.
If you’re in a right-wing social bubble - I don’t know what to recommend, my social bubble fulfills my left-wing content curation for me. But the overall effect is quite lovely - if both sides are complaining about something, it’s probably something like legitimate. If both sides are complaining about mirror images of something - it’s probably a legitimate but difficult problem that is being misrepresented to look simple. And if only one side is complaining about something - well, take it with a grain of salt.
In a sense, it’s only been relatively recently that we’ve actually entered the information age; the earlier golden age of the internet had many great things, and in some respects at least many fewer terrible things, but it was also so very small - and so very often wrong. Not in a “disinformation” way (that’s another topic entirely), but in a “this person doesn’t actually know what they’re talking about and nobody else I can find is even talking about this” kind of way.
Compared to today, the last golden age of the internet was downright claustrophobic.
If I seem bizarrely optimistic - well, I remember when I was telling everybody about this great new website called Google, which opened up huge new swathes of the internet to explore. (Its precursors were absolutely terrible, you see.) And I remember when I was telling everybody about this great new website called Yahoo, which opened up huge new swathes of the intertnet to explore. (Its precursors were also terrible, you see.) And if I’m not terribly impressed by Google now, with its increasingly claustrophic perspective - well, there twice was a new search engine that completely revolutionized the internet and rapidly replaced the archaic and claustrophobic “competitors”. Third time probably isn’t the charm, but things will get better for a while first.
I’m half-tempted to write a replacement myself, some days. P2P - peer to peer, which means it runs a little bit on everybody’s machine - offers a powerful decentralized platform which can’t easily be forced by foreign or domestic governments into compliance with their political agendas. There’s even already a P2P search engine - it just hasn’t really taken off yet, and it doesn’t do some things I think it needs to do to do so.
Centralization is, centrally, the problem that enabled the internet dark ages to occur at all; and while we are now increasingly rightfully suspicious of centralization (Google, Facebook, Reddit, Apple - basically every corporation in which some important facet of the internet has been invested, has used their position, often at the behest of government, to advance particular interests), there are still some heavily centralized areas.
A problem that I expect we’ll eventually need to face, and which offers room for a future third golden age, is - TCP/IP is probably built into your operating system, along with its sister protocol, UDP. That is, the protocol you use to use the internet has a central authority - the operating system - whose behavior is largely opaque to you. This hasn’t been an issue yet, to my knowledge - but it would surprise me if it doesn’t eventually become one, as governments seek ways to tame the internet. Certainly it would surprise me if there weren’t already surreptious logs being kept of flagged-suspicious TCP/IP requests. It has some flaws - the vulnerability to DDoS attacks being a key issue, albeit one that can be mostly dealt with, at least for now.
Going up a level, DNS servers; these are centralized authorities that, when you type “www.somewebsite.com”, get asked where this website actually “lives” in the internet. (That’s simplistic to the point of being mildly misleading, but, close enough). These have been regular vectors of attack for both governments and more regular bad actors; whether directly, like when the US government seized the domains of some pirating websites, or indirectly, as with China’s spoofing of DNS servers. There are alternatives to the default authorities; I couldn’t say whether they’re any good, it has yet to be a problem for me personally.
Then there is SSL, the protocol on top of TCP/IP that provides security (for now). You can at least write your own SSL protocol implementation - I’ve done it, and it isn’t any fun. However, the bigger issue is that SSL ultimately relies on Certificate Authorities - which, again, are centralized authorities that say whether or not somebody is who they say they are. These have been subject to government attacks, sometimes in conjunction with an attack on a DNS server; for example, a few years ago a Middle Eastern government (I cannot recall which one, it has been some time) spoofed a certificate authority to funnel certain kinds of communications through one of its own servers (which the spoofed certificate authority validated when asked), allowing it to identify protesters who, uh, were kind of relying on anonymity to avoid being murdered by said government.
Going up a level further, we start to get into what we actually think of as the internet; websites like Google. These are centralized only in that network advantages are immense. But there was a day that it looked like Yahoo couldn’t be toppled - and along came Google. MySpace gave way to Facebook. These companies, and their services, are still around - although I think maybe Yahoo outsourced its search engine capabilities to Microsoft. I still use its e-mail functionalities sometimes - I was amused one day to discover an e-mail from somebody, a couple of years prior, threatening me by showing that they had my password, a low-security password I used for a variety of different websites at the time. There never was any follow-up to that threat; I guess they realized it was an effectively dead account - mostly just accumulating spam and very rarely used to look for, say, a decade-old e-mail that contained a MIDI I once e-mailed somebody - and moved on.
Which is to say - their centralization isn’t really built on a solid foundation. Google will look indestructible, right up until it is destroyed, replaced with some other hot new search engine that doesn’t direct you to useless websites whenever you look up anything even vaguely medical. I recently wanted to see the statistics on sodium deficiency - see whether it has increased or not as people’s sodium intake has decreased. Sorry, no can do; here are ten official websites from respectable institutions telling you the symptoms of sodium deficiency, and then a series of websites whose utter irrelevancy seem deliberately designed to make me give up. A decade ago, I could type that in, and if I didn’t find it immediately, I could then switch to images - and I’d probably find a graph or two, and once I found those, I could drop into the websites and see what the authors had to say about it. Granted, I had to step through that process, which was kind of a hacky way of going about finding information, but at least it was available.
Now I’m relegated to Google Scholar, which, I notice, has been removed from the quick links under the search bar at some point, and which has, like all of Goog le’s search functionalities, gradually gotten increasingly useless.
Part of this isn’t strictly Google’s fault - SEO, that is, search engine optimization, has made this task increasingly difficult. But also - maybe it is entirely Google’s fault for relying so heavily on these algorithms in the first place.
Reddit, I think, suggests a path forward for search engines. Not Reddit, which is itself a corrupted shell of its former self - and I say that as somebody who never had an account. But rather, the way it self-organized; upvotes and downvotes. These, of course, have their own flaws and exploits - Urbit’s approach of requiring an up-front investment to participate I think has some merit to it, even if I don’t really care for Urbit’s design philosophy itself, which has pyramid scheme vibes to me, as it limits the ability of bad agents to use things like botnets to accomplish things. And there is, of course, the very central issue that truth isn’t a matter of popular opinion.
The trick to decentralization here is to mix in a little bit of centralization - just a tad. In a word, clustering. Imagine, for a moment, a “more like this” button, when searching, that elevates the prominance of upvotes for other websites based on whether the upvotes came from users who also upvoted that particular website - or a “less like this” button which suppresses the same. All operating on a decentralized P2P network, where each peer’s information is treated as something like a cluster of its own. That is - every user becomes a content curator. A return to an artisinal internet, but scaleable.
So a user searching for information on masks can prefer, or disprefer, users who upvoted the website that says that dewormers are an effective treatment for COVID-19.
If you’re horrified by that - bluntly, you’re part of the problem here. You want to control information. Remember, the alternative to your centralized authority with perverse incentives is just a -different- centralized authority with perverse incentives. “But they’ll be directed to websites which deny elections!” We’ve covered that, but also - what’s changed from the status quo? And when the next Google goes in a conservative direction, and starts suppressing the information that supports the causes you think matter - well, here’s the roadmap out of that maze.
I probably need to stop writing this at some point, and this looks like as good a point as any. I’m optimistic about the internet; in some significant ways, things look like they’re getting better; people are more suspicious of things they ought to be suspicious of, and in particular, they are increasingly suspicious of centralization itself, or at least its proxy agent corporatization. Or to frame that maybe slightly differently - the fact that so many people are increasingly pessimistic is a reason, in itself, for optimism. They’ve noticed the problem! Which means that, if you present them with a solution - maybe they’ll listen.
1. "I thought the problem was that we had no common enemy to unite against since the fall of the Soviet Union. But no, even the unity we possessed then was illusory, enforced by widespread censorship of the media."
I've speculated a bit about that, too. https://twitter.com/aronro/status/1547689913033768962
2. "upvotes and downvotes. These, of course, have their own flaws and exploits ..."
True. And Twitter's Community Notes (formerly "Birdwatch") has taken an interesting approach to mitigating at least some of the flaws with upvoting/downvoting – see the second paragraph and last sentence here, especially:
https://communitynotes.twitter.com/guide/en/contributing/diversity-of-perspectives.html
"Community Notes aims to identify notes that many people on Twitter will find helpful, including people with different points of view.
"To find notes that are helpful to the broadest possible set of people, Community Notes takes into account not only how many contributors rated a note as helpful or unhelpful, but also whether people who rated it seem to come from different perspectives.
"Community Notes assesses "different perspectives" entirely based on how people have rated notes in the past; Community Notes does not ask about or use any other information to do this (e.g. demographics like location, gender, or political affiliation, or data from Twitter such as follows or Tweets). This is based on the intuition that Contributors who tend to rate the same notes similarly are likely to have more similar perspectives while contributors who rate notes differently are likely to have different perspectives. If people who typically disagree in their ratings agree that a given note is helpful, it's probably a good indicator the note is helpful to people from different points of view."